AI experts at leading universities favor creating a federal "Department of AI" or a global regulator to govern artificial intelligence over leaving that to Congress, the White House or the private sector.

The big picture: That's the top-level finding of the new Axios-Generation Lab-Syracuse University AI Experts Survey of computer science professors from top U.S. research universities.

The survey found experts split over when or if AI will escape human control — but unified in a view that the emerging technologies must be regulated.

• "Regulation" was the top response when asked what action would move AI in a positive direction.
• Just 1 in 6 said AI shouldn't or can't be regulated. Only a handful trust the private sector to self-regulate.
• About 1 in 5 predicted AI will "definitely" stay in human control. The rest were split between those saying AI will "probably" or "definitely" get out of human control and those saying "probably not."

The intrigue: No one individual is highly trusted to deal with AI issues.

• President Biden took the top spot, with 9% of respondents — slightly higher than Sundar Pichai, Elon Musk or Sam Altman. Mark Zuckerberg and Donald Trump drew 2% and 1%, respectively.

Why it matters: The findings come ahead of a series of AI forums on Capitol Hill, beginning Sept. 13 and hosted by Senate Majority Leader Chuck Schumer, which will feature top U.S. tech CEOs and AI specialists.

• The insider perspectives reflected in the survey may help to inform policy, business and media perspectives on AI issues.

By the numbers: Respondents expressed greater worry about discrimination and bias resulting from AI (42%) than about the risk of mass unemployment (22%).

• 62% predicted AI will increase racial, gender and economic disparities.
• Customer service (77%); art, design or content creation (41%); and administrative and support services (39%) were the three sectors experts said were most likely to experience job losses in the next five years due to AI.
• Technology (62%), data and research analytics (44%) and health care (34%) were the three sectors experts predicted would see the most productivity growth.

Zoom in: A plurality (37%) supported having a federal agency regulate AI, compared with a global organization or treaty (22%), Congress (16%), the White House (4%) or the private sector (3%).

• The White House has so far published a blueprint for an AI bill of rights and obtained voluntary safety commitments from companies.

How it works: The survey includes responses from 213 professors of computer science at 65 of the top 100 U.S. computer science programs, as defined by SCImago Journal rankings.

• An experts survey does not necessarily reflect the views of the population at large. It is different from a poll, which looks at a random sample of 1,000 or more U.S. adults and carries an estimated margin of error.
• The computer science professors surveyed are not a representative sample of the wider population, and while experts' views may sometimes track with the general population they may differ if their views are shaped more by their understanding of technology than by expertise in politics, media or other realms.
• Experts from domains beyond computer science were not included in this survey, but they bring important perspectives to debates over AI as well.

What they're saying: Cyrus Beschloss, CEO of Generation Lab, said the most interesting finding is "definitely how much experts think we should regulate AI via a new Department of AI rather than leave it to Congress, POTUS or global charter."

• Beschloss noted that while doom and gloom permeates many AI discussions, "the threat of a human-crushing AI is not a concern to most of the smartest voices in the AI conversation."

Methodology: This Axios-Generation Lab-Syracuse University AI Experts Survey was conducted July 14-Aug. 6, 2023, with an online survey distributed by email.

• A listing of the participating institutions and additional details about the methodology may be found at the survey site.

Zoom announced Tuesday it is building a new AI assistant into its software as it aims to hang on to its spot as the leader in video conferencing, reports Ina.

Why it matters: Tech companies have been racing to add generative AI to their products, with the effort particularly intense around helping workers get more out of meetings.

Details: Zoom's AI Companion, as the product is being dubbed, combines new features with a couple that the company had previously been testing.

• The AI Companion will allow people catching up on meetings to watch highlights, read summaries and ask questions of a chatbot, among other features.
• The AI features will be free to all paid users, Zoom said.
• Zoom is relying on its own proprietary AI models, as well as Meta's Llama 2 and models from OpenAI and Anthropic.

Between the lines: Zoom attracted the wrong kind of attention last month amid concerns that the company was planning to use customer conversations to train AI systems. CEO Eric Yuan admitted the company made mistakes and Zoom clarified its terms of service.

• The company reiterated that it will not do so and added that the new AI features, like its initial ones, will be off by default and that account owners and IT administrators will choose whether to enable the features.

The big picture: Google and Microsoft have been adding AI features to their rival products, as has Otter, with the ability to transcribe and summarize meetings emerging as table stakes.

• Otter has its OtterPilot bot that can attend meetings on Zoom. Microsoft's Teams or Google Meet. Those in a meeting can ask questions about the meeting content in real-time or even choose to send the bot in their place.
• Google announced last week a suite of AI features for Google Workspace, including the ability to have its Duet AI bot attend and record meetings.

What's next: Starting next year, Zoom said it will introduce AI features that will allow people to prepare for meetings by tapping a conversational interface that can search through "content across the Zoom platform, such as prior meetings, chats, and select connected third-party apps."

• Humanoid robots — such as Apollo from Austin-based Apptronik — are starting to work side by side with people. (Axios)
• Big tech companies are hiring red team hackers to break their AI models. (Forbes)
• Tech companies added 12,643 workers across the U.S. in August, though the overall number of tech jobs fell by around 189,000 — pushing the tech unemployment rate up to 2.1%. (CompTIA)
• Elon Musk threatened to sue the ADL after blaming it for sagging ad sales at X (formerly Twitter). (Axios)

Remember Intel's Ottoman PC (1999)? Or Packard Bell's 1993 scheme to turn the letter "j" on your keyboard into a mouse pointer?

• Probably not. But reading about these and other hardware dead ends from the past four decades — assembled by Ernie Smith's Tedium newsletter — is a hoot.