[CANCELED] A Million Turing Tests per Second: Detecting bots and fraud in the time of LLMs

​​Unfortunately, we have to cancel the webinar at the last minute because the speaker's leadership did not permit the talk.​​​​ We apologize for any inconvenience this may cause and thank you for your understanding. We would be delighted to see you at our other webinars, the schedule for which can be found here.

​The Data Phoenix team invites you to our upcoming webinar, which will take place on June 27th at 10 a.m. PDT.

• ​​​​​​​Topic: A Million Turing Tests per Second: Detecting bots and fraud in the time of LLMs

• ​​​​​​​Speakers: Unmesh Kurup (Director of ML, Intuition Machines / hCaptcha)

• ​​​​​​​Participation: free (but you’ll be required to register)

​The Turing Test was originally supposed to involve a back and forth between human and bot with the bot passing the test when the human was unable to distinguish bot from another human. Today, millions of websites, endpoints, and merchants make this same determination, essentially performing a Turing Test, without the luxury of extended conversations. Every decision has to be made in a matter of milliseconds. To make matters worse, LLMs have supercharged this process by making it trivially easy to create and adapt bots aimed precisely at besting our defenses. In this talk, I provide a brief glimpse behind the scenes of the fascinating back and forth between consumers and bots, and how this game of cat and mouse continues to evolve in the time of LLMs. 

​​Key Highlights of the Webinar:

• ​What is bot and fraud detection and how does it work?

• ​What do bots do to try and break detection?

• ​How is this battle evolving with the use of LLMs?

• ​How do we detect and defeat LLM use?

​​​Speaker

​Unmesh Kurup is the Director of ML at Intuition Machines where he leads the team responsible for the research, development, deployment, and maintenance of all ML models that mitigate bot and fraud activity. Previously, he was senior manager at LG electronics where he was responsible for LG's work on ML in resource constrained settings. He is the author of over 50 papers in AI/ML and the co-creator of Auptimizer, an OSS library for hyperparameter optimization and model compression.