Peter Slattery, PhD

MIT FutureTech

Affiliate Researcher

Understanding trends in computing and artificial intelligence and trying to explore i) the risks from artificial intelligence, ii) their importance, and iii) how organizations are responding to these risks.

Visiting Researcher

Communication, strategy and research focused on predicting and harnessing advances in computing and artificial intelligence.

Massachusetts Institute of Technology

Visiting/Affiliate Researcher

Psychology Today

At Psychology Today, I write about the impact of artificial intelligence on our lives and societies.

Check out my articles here: https://www.psychologytoday.com/intl/contributors/peter-slattery-phd

Writer

The EA Behavioral Science Newsletter curates papers, forum posts, reports, podcasts, resources, funding opportunities, events, jobs and research profiles that are relevant to the effective altruism and behavioral science community.

Subscribe here: https://forms.gle/Yvb9QU2vdF73VSV57

The EA Behavioral Science Newsletter

Ready Research

Ready Research are a group of volunteers who conduct research to provide evidence-informed, rigorous and implementable recommendations for practitioners working on significant societal problems

As a research collaboration, we

- identify research areas where behaviour change or improved decision-making would do the most good

- coordinate and conduct rigorous applied research to discover how to change behaviour or improve decision-making, and

- amplify the reach and impact of behavioural evidence so that it can help support doing the most good with what we have.

If you're trying to decide which messaging will best encourage prospective donors to contribute financially to your charity, then you should read about our charitable donations review: https://docs.google.com/document/d/1osAwuO1J9L2z3PDGQn6UkFEA-4vs5WV9eJ5hAASnS7o/edit?usp=sharing

If you’re working in animal welfare or trying to encourage the public to eat less meat and other animal products, then you should read our animal products consumption review: https://osf.io/mcdsq/

If you’re trying to decide on the policies that will most reduce harm and support a sustainable recovery from COVID-19, you should read about our Survey of COVID-19 Responses to Understand Behaviour (SCRUB) project: https://www.readiresearch.org/research

Co-Founder

Monash University

Senior Communications Officer 

Research Fellow

BehaviourWorks Australia

Senior Communications Officer

Based within Monash University, BehaviourWorks Australia (BWA) is a research enterprise that brings leading behaviour change researchers and practitioners together to find behavioural solutions to real-world problems.

In BWA, I work on a wide range of projects (e.g., related to environment, climate, health, finance, pandemic response etc) in a wide range of roles (e.g., as project manager, research lead, reviewer, or report writer, or research translator etc).

In late 2017, I started as an embedded behavioural science researcher and consultant working with(in) two Sydney based NSW Government offices. Most of my work focused on behaviour change related to the environment. I learned a lot about how decisions and policies are made in government, and also that I lacked the background, patience and project management skills to thrive in these environments. 

In late 2018, I moved towards working on a wider range of topics, focusing more on rapid research synthesis and dissemination, and having greater autonomy to innovate and start new projects. This led to a range of reviews and publications, and services, such as the 'BWA research database', the scale up toolkit and better BWA 'hackathons'). 

In early 2020, I ideated and started the Survey of Covid-19 Responses to Understand Behaviour (SCRUB) project. More than 50,000 surveys in more than 40 countries were completed. The vast majority of this was Australian data, but we also collected over 3000 international responses. This work had a significant impact on government policy over the 15 months that it ran. 

Since late 2021, I mostly focus on improving our research communication and dissemination. This involved improving our processes, website, content, social media presence and analytics. This work led to BWA winning Behavioural Science organisation of the year in the Habit Weekly awards.

MIT Initiative on the Digital Economy

The University of Queensland

Developing a course on skills for doing good.

The SCRUB project

The Survey of Covid-19 Behaviours (SCRUB) aimed to provide current and future policy makers with actionable insights into public attitudes and behaviours relating to the COVID-19 pandemic. SCRUB was incubated as a READI project with international collaborators, and received pilot funding from the Monash Sustainable Development Institute at Monash University, and went on to be funded by the Victorian Government for a total of 21 study waves over 15 months. Findings from the project were an input to state policy decisions during the Australia COVID-19 pandemic. After each wave of data collection, we generated and disseminated a report for relevant policy makers. More than 50,000 surveys in more than 40 countries were completed. The vast majority of this was Australian data, but we also collected over 3000 international responses. SCRUB was mentioned in the media more than 200 times and published in several scholarly articles.

👋 I use research, communication and entrepreneurship to work on pressing societal problems.

My PhD examined how to use websites to encourage volunteering and philanthropy, but my research is varied and interdisciplinary. See my publications here: https://lnkd.in/gTARsgp

I have started and run several organisations. Most were successful, but not all. My startup 'Collaboreat' (later 'CompareMeals') failed (and made me very unhappy) but taught me many valuable lessons. My main side project at the moment is: https://www.readyresearch.org/ 

Despite hating public speaking, I have featured in the media more than 200 times. In 2020, I was the 16th most media engaged academic at Monash University. My media engagement ranges from articles for The Conversation, The Huffington Post and Psychology Today to TV interviews with channels such as the BBC and ABC News. You can listen to one of my interviews here: https://lnkd.in/gv4S-P2

I don't take myself too seriously or rate myself especially highly, and spend a lot of time on self-improvement. You can see (and copy) my 'annual review' or 'daily tracker' templates at:
https://bit.ly/3rOscb9
http://bit.ly/2MhHc0W

Outside work, I read, surf, swim and hike, relax in baths and hammocks (though never at the same time), and explore my (too) many interests. I recommend reading Behave by Robert Sapolsky and If This Is a Man/The Truce by Primo Levi, if you are looking for book recommendations. 

I am influenced by the ideas of effective altruism and 'doing good better' and care about my long term impacts on sentient creatures.

I try to help other people working on important issues where I can. Feel free to reach out if you would like my help, but please note that I am generally maxed out, and often can't make much, if any, time for new projects.

Apologies in advance if I don't respond to a message - I can't always handle the volume of messages that I get on LinkedIn and other social media (not without risking another burnout!). 

For work and collaboration enquiries, email: pslat@mit.edu

🌟 Career Updates: In March, I transitioned from a visiting position to become a Research Affiliate at MIT FutureTech. 

I have accepted a full-time position, which will probably start in September. I plan to leave Australia in May to go back to Ireland while I wait for my visa.

I have funding to work independently in the meantime. 

I recently finished working at BehaviourWorks Australia. I had a very good time there and developed a lot. I can highly recommend them to anyone who seeking a career, education or research services.

👋 Follow for AI, Behavioural Science, Systems Change and Self-Improvement | MIT FutureTech

🎉 I am very excited to see this book released! 

'Inspiring Change: How to influence behaviour for a better world' offers a step-by-step, evidence-based guide on how to develop your own behaviour change programs and interventions.

Each of the 12 chapters was written by different members of BehaviourWorks Australia's research team (including me at the time) based on extensive experience across diverse areas such as health, climate change, energy, water, waste, pollution, biodiversity, biosecurity, education, social inclusion, finance and safety.

The book features real-world case studies involving partnerships with Ambulance Victoria, the Transport Accident Commission, federal and state governments, the Environment Protection Authority and Sustainability Victoria.

It's applied, practical, insightful and well worth reading.

You can order here: https://lnkd.in/evuwsvQU

Thanks to Liam Smith, Jim Curtis, Peter Bragge and Paul Kellner for leading the project. Glad I could be involved. 

#BehaviourChange #BehavioralScience #sustainability #systemschange

📄 An interesting study on AI and cognitive bias.

This study examined the representativeness heuristic a cognitive bias where people judge the likelihood of an event based on how closely it resembles a typical example, rather than considering broader facts or statistical evidence.

Experiments on four LLMs (GPT-3.5, GPT-4, PaLM 2, and LLaMA 2) found that LLMs have demonstrated representativeness heuristic biases similar to those observed in humans, with accuracy ranging from 10% to 50% across various tasks.

AI psychology is going to be a very important area of future research and well worth considering if you are planning a research project!

Follow MIT FutureTech if you are interested in technological trends and their social implications.

Paper

A fascinating example of how AI may change our societies and behaviors. 

Just as the mediation of social media homogenized our culture, the mediation of LLMs may homogenize our vocabulary and communication norms. 

I look forward to delving deeper into the topic in future posts :)

Follow MIT FutureTech if you are interested in technological trends and their social implications.

Professor of Strategy, Copenhagen Business School

Ha! I've noticed that *lots* of LinkedIn posters have suddenly started using the annoying verb "delve" ("delve into" etc.). For some reason, ChatGPT loves "delve" (2023 was the first full year of ChatGPT). Stop using ChatGPT for writing your LinkedIn posts!  ;-)  (H/T https://shorturl.at/LMX58). 

📄 Why are large language models (LLMs) persuasive? 

"Large Language Models (LLMs) are already as persuasive as humans.
However, we know very little about why. This paper investigates the persuasion strategies of LLMs, comparing them with human-generated arguments."

"The study reveals that LLMs produce arguments that require higher cognitive effort, exhibiting more complex grammatical and lexical structures than human counterparts. Additionally, LLMs demonstrate a significant propensity to engage more deeply with moral language, utilizing both positive and negative moral foundations more frequently than humans."

"These results are particularly significant in the context of the potential dark side of LLMs in general, but also for discussions on misinformation and AI ethics in particular."

"The finding that LLMs can employ moral-emotional language with persuasive effectiveness has implications for how AI systems could amplify social biases and moral divides."

"This raises ethical questions about the role of AI in shaping societal norms and values, and the responsibilities of AI developers to address and correct for potential biases within LLMs."

Follow MIT FutureTech if you are interested in technological trends and their social implications.

"OpenAI recently changed their terms of service to no longer prohibit military and warfare use cases"

The military are already testing AI and thinking about how to implement it in the future.

This is understandable but worrying. No-one wants to do anything risky here, but no-one wants to be left behind.

What happens is sometimes called race dynamics - a coordination problem where perhaps all militaries would like to slow down, if they believed they were on parity or ahead of their competition, but no-one does as no-one believes they can afford to do so.

We have seen lots of companies releasing AI that they clearly didn't think was bad, then realizing it had some major issue. Behind this was a commercial pressure to get something out.

If companies are rushing out unsafe AI for commercial reasons, then how might we expect militaries to act when lives are being lost in wars? 

If China has superior AI on a battlefield with the US, will anyone in the US Military say - 'let's just lose this war rather than put our superior AI in charge so we win'? Unlikely. If they do, will anyone listen. 

What happens when that AI acts in ways we didn't expect, like nearly all the advanced AIs we have, but now with advanced weaponry?

I think that there is a non-negligable chance this happens, or something comparably serious. Hopefully history will prove me to be overcautious, but until it does, I really think we need to think ahead about all these risks. 

We have time to act now, but we may not later. AI is here, better AI is coming, and soon, in some timeframe, our lives will be changed forever for much better or much worse.

Follow MIT FutureTech if you are interested in technological trends and their social implications.

AI models chose violence and escalated to nuclear strikes in simulated wargames

Open article: AI models preferred to launch nuclear strikes in wargames test by euronews.com • 4 min read

AI models preferred to launch nuclear strikes in wargames test

Are you interested in promoting prosocial behavior, better decision-making, or improved moral reasoning? See the latest research on these topics and much more in the recent April (#12) edition of the Effective Altruism Behavioral Science Newsletter.

This quarter, Timothy Caballero and I curated:
📖 6 publications,
📝 11 preprints & articles,
💬 6 forum posts,
🎧/🎦 5 podcasts & videos, and;
💼 3 new opportunities

🔔 Subscribe to receive future editions of the newsletter here: https://lnkd.in/g3MZZEUd

Sign up to Habit Weekly for weekly insights into behavioral science and design.

#behavioraleconomics #animalwelfare #fundraising #effectivealtruism
#psychology

Open article: 💌 The EA Behavioral Science Newsletter, April 2024 (#12) by preview.mailerlite.com

💌 The EA Behavioral Science Newsletter, April 2024 (#12)

Reflections from Yoshua Bengio - the most cited computer scientist in the world: "Instead of decades to centuries, I now see [human-level AI] as 5 to 20 years with 90% confidence."

"I started reading more about AI safety and came to a critically important conclusion: we do not yet know how to make an AI agent controllable and thus guarantee the safety of humanity! And yet we are – myself included until now – racing ahead towards building such systems."

"It is difficult to digest such reflections and carry out the mindset shift that it entails. It is difficult because accepting the logical conclusions that follow means questioning our own role, the value of our work, our own sense of value."

"Being able to change one's views when faced with new evidence or new arguments is essential for the advancement of science, as well as to steer society towards a beneficial future. The more curious and interested we are in our mistakes, the more we learn, grow, evolve and broaden our capacity to impact others and the world in a positive manner."

"Unfortunately, to settle the AI debate, we can’t rely on mathematical models of how research, technology and politics are likely to evolve in the next decade under different interventions. We don’t have past experience interacting with machines more intellectually capable than us and thus obtain statistics of what is safe and what isn’t. Yet we cannot wait until irreversible damage is done in order to change course. "

"As scientists, we should avoid making claims we can't support; but as decision-makers we also ought to act under uncertainty to take precautions. In spite of our differences in points of view, it's time for our field of AI to seriously discuss the questions: what if we succeed? What if potentially dangerous superhuman AI capabilities are developed sooner than expected?"

💬 What do you think? 
Is Yoshua right to be concerned about these issues? 

📚 References/further reading
https://lnkd.in/dPApG4xH

Follow MIT FutureTech if you are interested in technological trends and their social implications.

#AI #technology 

Open article: Personal and Psychological Dimensions of AI Researchers Confronting AI Catastrophic Risks - Yoshua Bengio by yoshuabengio.org • 12 min read

Personal and Psychological Dimensions of AI Researchers Confronting AI Catastrophic Risks - Yoshua Bengio

🌟 These fascinating graphs from The Economist's article 'Societies change their minds faster than people do' got me thinking.

They show how generational replacement shifts public opinion.

One reason why generational replacement shifts public opinion is probably because each generation is exposed to an increasing breadth and diversity of information.

My generation had mobile phones and basic personal computers, but most of our information was from TV and books.

The current generation of kids are using laptops, mobile phones, and the internet in school and have access to vastly more data from a wider range of sources (which definitely has major downsides, in some cases).

The next generation might have advanced AI, VR worlds, robot assistants, Generative AI companions, and drones. 

It's crazy to think about how different that experience will be.

It's one thing to know about people in other countries via television or the internet, and another to be able to engage with them in VR.

Similarly, websites are much less effective at education than even current LLMs. 

I suspect that one impact will be that the next generation are much more concerned about and aware of the risks from AI, including risks to AI, because many will have immediate immersion in the technology and experience with AI companions and agents from a young age.

What do you think?

Follow MIT FutureTech if you are interested in technological trends and their social implications.

🤖 There's been a lot of debate whether AI will be a transformative technology that accelerates economic growth.

This is an excellent summary of the arguments for we should be cautious in our expectations.

They make three main points:

 1️⃣ The transformative potential of AI is constrained by its hardest problems

"Automating most tasks has very different effects on growth than automating all tasks: ...economic growth may be constrained not by what we do well but rather by what is essential and yet hard to improve... When applied to a model in which AI automates the production of ideas, these same considerations can prevent explosive growth."

In “What if we could automate invention,” Matt Clancy provides a final dose of intuition: Invention has started to resemble a class project where each student is responsible for a different part of the project and the teacher won’t let anyone leave until everyone is done... if we cannot automate everything, then the results are quite different. We don’t get acceleration at merely a slower rate—we get no acceleration at all."

2️⃣ Despite rapid progress in some AI subfields, major technical hurdles remain

"Steven Pinker's 1994 observation remain relevant: "The main lesson of thirty-five years of AI research is that the hard problems are easy and the easy problems are hard." The hardest "easy" problems, like tying one's shoelaces, remain. Do breakthroughs in robotics easily follow those in generative modeling? That OpenAI disbanded its robotics team is not a strong signal."


3️⃣ Even if technical AI progress continues, social and economic hurdles may limit its impact

"AI may not be able to automate precisely the sectors most in need of automation. We already “know” how to overcome many major constraints to growth, and have the technology to do so. Yet social and political barriers slow down technology adoption, and sometimes halt it entirely. The same could happen with AI."

"GDP is a made-up measure of how much some humans value what others produce, a big chunk of which involves doing social things amongst each other." 

💬 What do you think? Do you find the arguments convincing? How do you expect AI to ultimately reshape the economy and society?

📚 Resources
Essay: https://lnkd.in/eic2vW-S

For a more optimistic take see: Explosive Growth from AI: A Review of the Arguments from Epoch
https://lnkd.in/e_YaAYNw

Follow MIT FutureTech if you are interested in technological trends and their social implications.

#Ai #Economics

Open article: Why transformative artificial intelligence is really, really hard to achieve by thegradient.pub • 20 min read

Why transformative artificial intelligence is really, really hard to achieve

🔍How do humans spend the average day? I have often wondered what the average person does, and now I finally have more insight!

"Here, we estimate how all humans spend their time using a generalized, physical outcome–based categorization that facilitates the integration of data from hundreds of diverse datasets." 

"Our compilation shows that most waking hours are spent on activities intended to achieve direct outcomes for human minds and bodies (9.4 h/d), while 3.4 h/d are spent modifying our inhabited environments and the world beyond. The remaining 2.1 h/d are devoted to organizing social processes and transportation" 

"Globally, the time spent directly extracting materials and energy from the Earth system is small, on the order of 5 min per average human day, while the time directly dealing with waste is on the order of 1 min per day, suggesting a large potential scope to modify the allocation of time to these activities. "

"Our results provide a baseline quantification of the temporal composition of global human life that can be expanded and applied to multiple fields of research."

Fascinating. I hope they track this over time to see how it changes alongside new technologies. They should also really track time on social media and engaging with technology. What else?

Follow MIT FutureTech if you are interested in technological trends and their social implications.


📚 References/further reading
https://lnkd.in/eWDaTbxC 

Open article: The global human day | Proceedings of the National Academy of Sciences by pnas.org • 2 min read

The global human day | Proceedings of the National Academy of Sciences

Interesting - Spencer Greenberg tested the personalities of the major AI chatbots. Key findings:

Three out of four AIs identified as Enneagram type 7 (the "enthusiast"), drawing parallels to the Genie from Aladdin.

In the MBTI-style framework, ChatGPT and Copilot were ENTJs, while Gemini and Claude were ISTPs. All four AIs scored as T (thinking) rather than F (feeling).

Big Five results varied, with Copilot and ChatGPT scoring highest on openness and conscientiousness, and Gemini and Claude scoring lowest.

The AIs chose their own gender presentation, with Gemini and Copilot selecting feminine and ChatGPT and Claude choosing masculine.

I think that understanding AI models/agents is only going to become more important and relevant, and I'd love to see more work in this area!

Follow MIT FutureTech if you are interested in technological trends and their social implications.

a mathematician and entrepreneur focusing on improving society through social science

I personality-tested the major A.I. chatbots (Google Gemini, Microsoft Copilot, OpenAI ChatGPT, Anthropic Claude) using our "Ultimate Personality Test."
If you'd like to get an infographic for your personality, the link is here (it's free): https://lnkd.in/efjhRuwp

Some observations:

1) Masculinity/Feminity

I allowed the A.I.'s to pick their own gender presentation for the report. Gemini and Copilot chose feminine, and ChatGPT and Claude chose masculine (hence why you see the images you do in their results).

2) Enneagram results

Three of the four A.I.'s were Enneagram type 7 (the "enthusiast"), which is generally considered to be the enneagram of the Genie from the movie Alladin. Interestingly enough, both the Genie and these A.I.s are constrained intelligences whose role is to do the bidding of anyone who makes requests of them. I actually expected the A.I.'s to get type 5 (the "investigator"), which is generally considered to be the Enneagram type of Sherlock Holmes, but only Microsoft's Copilot (previously called "Bing") was assigned type 5.

3) Myers-Briggs-style results

Two of the A.I.'s were ENTJs in the MBTI-style/Jungian framework (OpenAI's ChatGPT and Microsoft's Copilot). In contrast, two were close to the opposite, getting ISTP (Google's Gemini and and Anthropic's Claude). All four A.I.s were scored as T (thinking) rather than F (feeling), which makes sense given their focus on facts rather than feelings.

4) Big Five results

In the Big Five personality framework (which is the one most used in academia and which has the most scientific backing), there was quite a bit of divergence across the systems. See the attached image for a summary of how each A.I. system scored on the Big Five:
https://lnkd.in/enHAiCC2

For each trait is shown on a scale from (1), which is low, to (5), which is high. 

All of the systems were low in neuroticism (1) except for Copilot, which had a middling level (3). 

All of the systems were in the low to middle range in extroversion, getting a (1) to (3).

None of the A.I.s had an extreme on agreeableness, with scores ranging from (2) to (4). 

Scores on openness and conscientiousness were more varied, with Copilot and ChatGPT being the highest on both and Gemini and Claude being the lowest on both.

5) Limitations
It is absolutely the case that if I ran this again or used a somewhat different prompt, I may have gotten somewhat different results. So, these shouldn't be taken as definitive results for any of these models. A more rigorous version of this would involve testing each model multiple times and exploring a variety of different prompts.

Here is the prompt I created and used for personality testing all of these A.I.'s:
https://lnkd.in/eGPnj_du

A big hat tip to Michael Nielsen and @AISafetyMemes for giving me the idea

💡 AI Outperforms Human Psychologists in Social Intelligence Test

A recent study published in Frontiers in Psychology compared the social intelligence of advanced AI language models like ChatGPT-4, Google Bard, and Bing to human psychologists. 

📄"The Human participants were a sample of male psychologists in the Kingdom of Saudi Arabia with one of two levels of education (Bachelor’s and doctoral students) at King Khalid University during 2023–2024. The study sample consisted of 180 participants, including 72 bachelor’s students and 108 doctoral students in counseling psychological program."

"We conducted a single evaluation for each AI model on August 1, 2023 of its SI performance using the Social Intelligence Scale (Sufyan, 1998). In each evaluation, we provided AI the same 64 standard SI scenarios. A link to the questionnaire was sent to human participants via e-mail. While the large linguistic models of AI were asked to answer the scale items individually and their answers were collected in a separate external file by directing a question to the AI models to choose the appropriate answer from the alternative points for each item in the scale."

📊 Key findings:
1️⃣ ChatGPT-4 outperformed all participating psychologists on the Social Intelligence Scale.
2️⃣ Bing surpassed more than half of the psychologists, especially those at the bachelor's level.
3️⃣ Google Bard's performance was comparable to bachelor's level psychologists but significantly lower than doctoral students.

💬 Does this surprise you? I didn't expect AI would beat humans at social intelligence.

What areas of intelligence do you expect Humans to outcompete AI in over the longer term? 

📚 References/further reading
Sufyan, N. S., Fadhel, F. H., Alkhathami, S. S., & Mukhadi, J. Y. A. (2024). Artificial intelligence and social intelligence: preliminary comparison study between AI models and psychologists. Frontiers in Psychology, 15.

Follow MIT FutureTech if you are interested in technological trends and their social implications.

#AI #technology #economics

🤖💼?: Whether AI automation will take all our jobs is becoming an increasingly common question.

Epoch have a pretty comprehensive report which explores and explains the approaches used to answer that question (linked in comments).

As you might expect, they find significant differences between approaches, both in terms of their methodologies and the predictions they make.

Key highlights:
📋 Approaches can be grouped into three main categories: task feature analysis, task-patent mapping, and automation forecasting surveys

🔍 Most research focuses narrowly on the technical potential to automate tasks, overlooking crucial deployment practicalities

⏱️ Most research offers ratings of relative job automaticity rather than a clear timeline for when automation happens.

MIT FutureTech have a recent paper on this which (I think) goes a little further than some of the previous literature in terms of the precision of the model and prediction - see link in the comments.

#AI #economics #jobs

This is just out from Boston Dynamics.

The combination of advanced robotics and AI is going to be very interesting.

We may soon see robots becoming much more autonomous and capable of performing complex tasks that were previously limited to humans.

Also being used regularly in environments that are hazardous for humans, like disaster sites, industrial settings, mining or warfare.

It's well worth paying attention to!

Follow MIT FutureTech if you are interested in technological trends and their social implications.

What Drives Progress in AI? 

In this new article, Tal and I explore the pivotal role data plays in driving advancements in artificial intelligence: from machine learning algorithms to Generative AI, the quality and quantity of data fueling these technologies determine their efficacy and impact. 

Follow MIT FutureTech if you are interested in technological trends and their social implications.

#AI #DataScience #MITFutureTech #Innovation #data #GenAI #artificialintellgience

Open article: What drives progress in AI? Trends in Data by futuretech.mit.edu • 4 min read

What drives progress in AI? Trends in Data

A short overview of some of the reasons to worry about the risks from AI from 80,000 Hours. Worth a watch if you haven't had a chance to think about these issues - AI is becoming a big deal, as you might have noticed!

0:00 Introduction
0:46 Existential risk from AI
02:45 AI systems are “black boxes”
03:26 AI systems might develop secondary goals
05:31 How could an AI disempower humanity?
06:13 Maybe it’ll be fine
06:51 Misuse of AI
07:18 The future is uncertain
08:16 Solutions
09:09 Conclusion

You can read the full article on risks from advanced AI here: 80k.info/ai-profile-video 

Full article: https://80k.info/ai-profile-video This is our first time producing this kind of video, so we’d really love to hear your feedback! Tell us what you thought of the video here: https://forms.gle/T9JGq5NA8Rrgf7NV6 Or, if you’d like to read...

Open video: Could AI wipe out humanity? | Most pressing problems on youtube.com

Could AI wipe out humanity? | Most pressing problems

🤔 Elon Musk predicts AI will overtake human intelligence next year, but what do the experts think?

🔍 Katja Grace and AI impacts have done the most comprehensive work to aggregate expert opinion on how AI will perform relative to humans (that I know of)

"In a new survey of 2,778 AI experts, experts gave a median 5%
chance that AI would cause human extinction.

The survey also found that powerful AI may arrive much sooner than many people expect. Experts now see a 50% chance of AI models outperforming humans in every task by 2047, assuming no major disruption to scientific activity — just over 20 years from now. Notably, this estimated date is 13 years sooner than the date arrived at by a similar survey in 2022, in which respondents estimated such AI would not arrive until 2060.

...

Experts expect systems will be able to do many tasks long before then, though. They thought there was a 50% chance that AI models would be able to produce a new song indistinguishable from a Taylor Swift song by 2027, and a 50% chance AI could write NYT best-selling fiction by 2031. They also predicted a 50% chance that AI models would be able to perform as well as the best humans in the prestigious Putnam math competition within eight years."

💬 What do you think? Do you agree with the experts or Musk?

📚 References/further reading
https://lnkd.in/eWnhh4Zx

Follow MIT FutureTech if you are interested in technological trends and their social implications.

“My guess is that we’ll have AI that is smarter than any one human probably around the end of next year.” “Within the next five years, the capabilities of AI will probably exceed that of all humans, Musk predicted on Monday during an interview.
 
“The capability of new artificial intelligence models will surpass human intelligence by the end of next year, so long as the supply of electricity and hardware can satisfy the demands of the increasingly powerful technology, according to Elon Musk.”
 
Musk has been consistently bullish on the development of not only artificial general intelligence, AI tools so powerful they can beat the most capable individuals in any domain, but almost everything from Mars to robots and skull implants. He predicted in 2019 there would be a million robo-taxis by 2020. In 2016, he said about Mars, "If things go according to plan, we should be able to launch people probably in 2024 with arrival in 2025." 
 
Both achievements are years away, and some think they will no longer happen. Most cities continue to resist self-driving vehicles even as some state governments support them. As for Mars, Musk has rolled back his plans with the latest prediction for a crewed mission being in 2029.
 
Yet he is becoming more aggressive about AGI. “Last year, he predicted full AGI would be achieved by 2029.” Now he has said 2025. I wonder why? Perhaps because he thinks he is losing control of the narrative to other tech bros, from Sam Altman to Marc Andreesson. They are making their own predictions, sometimes even outdoing Elon Musk. 
 
Another contradiction is that “Musk’s eagerness to join the AI race runs counter to the high-profile stance he took last year calling for a pause in advanced AI development. He warned of the “profound risks to society and humanity” posed by super-powerful AI tools and called for an immediate pause on training any system more powerful than OpenAI’s GPT-4, the market leading model.”
 
My take: the tech bros want to control the narrative. None of what they say has to come true because the media will consider their comments news items worthy of being repeated. The same holds for the big consulting companies such as McKinsey and Accenture. 
 
These tech bros and consulting companies announce predictions to keep their names in the news because they know that this drives big contracts with CEOs, politicians and other big honchos who play the same game. It doesn’t matter whether their predictions come true or that anything is accomplished. It just matters that big honchos keep their names in the news and appear relevant to other big honchos. Whether engineers and scientists believe the #hype or not doesn’t matter, the big honchos are the ones Musk is talking to, and the media is happy to play along.
#technology #innovation #startups #artificialintelligence

Open article: Elon Musk predicts AI will overtake human intelligence next year by ft.com • 3 min read

Elon Musk predicts AI will overtake human intelligence next year

This is very impressive, particularly the speed and accuracy of the movement. Robots have already surpassed my performance in the kitchen!

Hopefully these videos aren't overly rehearsed.

LinkedIn 'Top Voice' in Technology & Innovation | Group Director @ Meta [Facebook, Instagram, WhatsApp, Threads, Messenger, Quest, and more...]

The robots are coming. 🔥

If you’re not yet convinced that general purpose robot helpers will be normal in the not too distant future, I highly recommend this short demo.

It’s called ‘Astribot’, and it was just released by a Chinese research company. A lot is still unknown about their work, but the video really brings to life the incredible potential as robotic assistants come closer to the tipping points required to take off.

The race to develop general purpose robots has been heating up for years, but still with major limitations on speed, dexterity, and complexity of tasks.

… but this is changing.

The continued convergence of advanced robotics, computer vision, and AIs with multi-model understanding will undoubtedly unlock a new global industry.

Not if. When.

Amazon already has close to a million robotic unit deployed in their warehouses (up from zero just 12 years ago!), and a simple use case like robotic lawn mowers is already a multi billion-dollar industry on its own.

Imagine when the industry unlocks true general purpose at a reasonable price point.

How quickly will it then become normal to have a robot helper in your house?

It’s still early days, with many hurdles still to overcome.

… but they’re coming.

Fascinating future ahead.

Do guilt appeals work? An interesting meta-analysis of 127 effect sizes from 26 studies with a total of 7,512 participants.

Findings
The overall effect size of guilt appeals was small, with Hedges' g = 0.19, 95% CI (0.10, 0.28). 

Guilt appeals that 

- attributed responsibilities to the perceiver produced smaller effects compared to those not attributing responsibilities.

- depicted controllable causes had smaller effects than those not depicting controllable causes.

- attribute a victim's suffering to an unstable cause had greater effects on outcomes than did those not depicting such causes.

- recommended reparative behaviors had significantly stronger effects

- were employed in environmental advocacy and education contexts were more effective than those in advertising/marketing contexts.

Open article: Frontiers | When guilt works: a comprehensive meta-analysis of guilt appeals by frontiersin.org • 60 min read

Frontiers | When guilt works: a comprehensive meta-analysis of guilt appeals

Taplio

Peter Slattery, PhD's Linkedin Analytics