Seems like a change in tone from past interviews and posts. Is it because he truly believes this, or because he needs to tone things down for PR, internal business, or government regulation reasons?
REGULATION
-
Stanford Study Reveals Legal AI Errors Deepen Inequalities
By
–
A new study from HAI & Stanford RegLab shows pervasive errors in legal AI with a concerning implication: “Our findings suggest that the current limitations of LLMs pose a risk of further deepening existing legal inequalities.” (via @thehill
) -
Google Circle To Search Privacy Data Collection AI Models
By
–
Google’s VP of Search Cathy Edwards just answered that question on stage at Unpacked. She said Google will only be able to see whatever a user highlights via Circle To Search. I also wonder how the new AI tools will use device owners' data to improve text & image AI models, etc.
-
Samsung Circle To Search: New AI Feature Raises Privacy Concerns
By
–
At #SamsungUnpacked, Samsung and Google just announced a "Circle To Search" tool that lets users search anything on their app by circling, highlighting or tapping on anything without switching apps. A big question I have: What do all these new AI features mean for user privacy?
-
Assign Roles to Large Language Models
By
–
10. Assign a role to the large language models. pic.twitter.com/1od0vv86Be
— The Rundown AI (@TheRundownAI) 17 janvier 202410. Assign a role to the large language models.
-
AI Political Use Policy Guidelines and Considerations
By
–
Not surprising. What’s your policy for political use?
-
EU Digital Skills Target Increases to 80 Percent by 2030
By
–
Shares of digitally skilled people in Europe: – In 2023, 56% of people in the EU aged 16 to 74 had at least basic overall digital skills
– By 2030, the target is to increase it to 80%. More> https://
bit.ly/3Slf5hb @DigitalEU via @antgrasso #DigitalEUAmbassador #FutureofEurope -
OpenAI Internal GPT-5 Access: Competitive Advantage Concerns
By
–
And I suppose the same holds true for GPT-5. Whenever it is done training, and moves into red teaming, does OpenAI keep a raw internal version to use for themselves and their own business? If so, what capabilities and powers does that give them that the rest of us won’t have?
-
Hidden AI Usage in US Elections: Persuasion and Personalization
By
–
I think about this post a lot in relation to the current election cycle in the U.S. AI usage by politicians will be obvious at times, but it’s the hidden uses where AI may have the greatest impact, such as leveraging AI’s persuasive abilities in messaging and personalization.
-
Government AI Rules Risk Jeopardizing Modernization Efforts
By
–
New government rules around AI could put much-needed modernization efforts in jeopardy, say HAI senior fellow Daniel E. Ho and @UMichLaw
’s @nicholas_bagley
.