Kobi Hackenburg

phd candidate @oiioxford @uniofoxford | research scientist @AISecurityInst | AI, social data science, persuasion with language models

[Kobi Hackenburg @KobiHackenburg]

🚨 New today in @ScienceMagazine !!🚨

We’re publishing the results of the largest AI persuasion experiments to date: 76k participants, 19  LLMs, 707 political issues

We examine “levers” of AI persuasion: model scale, post-training, prompting, personalization, & more…

🧵:

[Kobi Hackenburg @KobiHackenburg]

Very excited to see this amazing work by @lujainmibrahim out today in @Nature :)

[Lujain Ibrahim @lujainmibrahim]

🚨Very excited to see our work on warmth & sycophancy in LLMs out in @Nature today!🚨

We study what happens when LLMs are fine-tuned to be warmer, and find that warmth and sycophancy can be linked, with warm models showing higher errors on a range of benchmarks (🔗s below)

[Paul Röttger @paul_rottger]

New paper w/ @AISecurityInst: AI writing assistance distorts how others perceive AI users and their opinions.

Millions of people now use AI to help them write and communicate. In three large experiments (14k participants, 3m+ human ratings) we show that AI writing assistance systematically distorts writer personas – their perceived beliefs, personality, and identity. These distortions are consistent across AI models and persist even under realistic conditions of human oversight.

🧵

[Kobi Hackenburg @KobiHackenburg]

In other words, we measure distortions between purely human-authored writing, and *human edited*, AI-assisted writing *which humans preferred to their own original writing*

Has been great to work on this with @paul_rottger @hannahrosekirk @summerfieldlab. Feedback very welcome!

[Kobi Hackenburg @KobiHackenburg]

Very excited to see this out!

We had a hunch that pervasive use of AI writing assistance for political opinion expression must be ~doing something~ to how those opinions are perceived in aggregate

In large RCTs, we use a nifty within-subjects design to show exactly what :)

[Paul Röttger @paul_rottger]

New paper w/ @AISecurityInst: AI writing assistance distorts how others perceive AI users and their opinions.

Millions of people now use AI to help them write and communicate. In three large experiments (14k participants, 3m+ human ratings) we show that AI writing assistance systematically distorts writer personas – their perceived beliefs, personality, and identity. These distortions are consistent across AI models and persist even under realistic conditions of human oversight.

🧵

[Kobi Hackenburg @KobiHackenburg]

By distortion, we mean the difference in how third-party readers (blind to authorship) perceive a writer's own text vs. their AI-assisted text.

Our design mimics the real world, where users can freely edit AI outputs and are free to *not use* AI-assisted outputs they don't like

[Kobi Hackenburg @KobiHackenburg]

@j_kalla @Ben_Tappin @lukebeehewitt @hauselin @realmeatyhuman @EdSaunders @CatherineFist @HelenMargetts @DG_Rand @summerfieldlab @AISecurityInst You can read the full paper in @ScienceMagazine here:

https://t.co/Xy94gQmizF

Supplementary materials can be found here:

https://t.co/sWXFPhOE0u

[Kobi Hackenburg @KobiHackenburg]

🚨 New today in @ScienceMagazine !!🚨

We’re publishing the results of the largest AI persuasion experiments to date: 76k participants, 19  LLMs, 707 political issues

We examine “levers” of AI persuasion: model scale, post-training, prompting, personalization, & more…

🧵:

[Kobi Hackenburg @KobiHackenburg]

@j_kalla @Ben_Tappin @lukebeehewitt @hauselin @realmeatyhuman @EdSaunders @CatherineFist @HelenMargetts @DG_Rand @summerfieldlab I’m also very grateful to many more people @AISecurityInst for making this work possible! There will be lots more where this came from over the next few months 💪