Learning Library

← Back to Library

AI Boosts Fantasy Football, Fuels Cybercrime

Key Points

  • IBM’s Watson X powers new AI features in the ESPN Fantasy Football app, delivering millions of insights—including waiver‑grade and trade‑grade scores—to help roughly 11 million managers make smarter roster moves.
  • The AI models ingest and analyze vast amounts of news, expert opinion, and injury reports, generating up to 48 billion data points to personalize player recommendations and evaluate trade value.
  • While generative AI chatbots such as ChatGPT have surged in popularity, cyber criminals quickly adapted these tools to craft convincing phishing emails, prompting AI providers to add guardrails against misuse.
  • Malicious actors are now deploying specialized AI models like WormGPT (built on the GPT‑J architecture) to automate business‑email‑compromise attacks, making fraudulent messages harder to detect by traditional cues such as grammar errors.
  • The episode highlights both the beneficial impact of AI‑driven analytics in fantasy sports and the emerging security challenges posed by AI‑enabled cyber‑crime.

Full Transcript

# AI Boosts Fantasy Football, Fuels Cybercrime **Source:** [https://www.youtube.com/watch?v=pDkIEHeHD-A](https://www.youtube.com/watch?v=pDkIEHeHD-A) **Duration:** 00:04:11 ## Summary - IBM’s Watson X powers new AI features in the ESPN Fantasy Football app, delivering millions of insights—including waiver‑grade and trade‑grade scores—to help roughly 11 million managers make smarter roster moves. - The AI models ingest and analyze vast amounts of news, expert opinion, and injury reports, generating up to 48 billion data points to personalize player recommendations and evaluate trade value. - While generative AI chatbots such as ChatGPT have surged in popularity, cyber criminals quickly adapted these tools to craft convincing phishing emails, prompting AI providers to add guardrails against misuse. - Malicious actors are now deploying specialized AI models like WormGPT (built on the GPT‑J architecture) to automate business‑email‑compromise attacks, making fraudulent messages harder to detect by traditional cues such as grammar errors. - The episode highlights both the beneficial impact of AI‑driven analytics in fantasy sports and the emerging security challenges posed by AI‑enabled cyber‑crime. ## Sections - [00:00:00](https://www.youtube.com/watch?v=pDkIEHeHD-A&t=0s) **AI Powers Fantasy Football Insights** - IBM’s Watson X fuels AI-driven waiver and trade grades in ESPN’s Fantasy Football app, delivering billions of player insights and recommendations to millions of managers. ## Full Transcript
0:00Ai and fantasy football and a look at 0:03generative AI tools being used by 0:05criminals all on this episode of IBM 0:07Tech now what's up y'all my name is Ian 0:10and I am back to bring you the latest 0:12and greatest news and announcements 0:14about IBM 0:15technology the NFL season is in full 0:17swing and if you're one of the estimated 0:1960 million Americans who play Fantasy 0:22Football you're likely spending a lot of 0:24time pouring over stats and stories to 0:26find that edge and if you're one of the 0:28many like me whose fantasy team is 0:30struggling you could probably use a 0:32little help luckily for you IBM partners 0:35with ESPN Fantasy Football to use AI 0:38models built with Watson X to provide 11 0:41million fantasy managers with a data AI 0:44infused experience that transcends 0:46traditional 0:47statistics the AI models built with 0:49Watson X ingest and analyze millions of 0:52news stories opinion Pieces by fantasy 0:54experts and reports on player injuries 0:57the models are expected to produce more 0:59than 40 48 billion insights for Fantasy 1:01managers this year including 1:03recommending trade opportunities and 1:05identifying waiver wire players two key 1:08types of player insights are now 1:10integrated into the ESPN Fantasy 1:12Football app waiver grades and trade 1:15grades using neural networks and 1:17advanced natural language processing 1:19waiver grades give a personalized rating 1:21for the value a player would add to your 1:23team this nuanced algorithm delves deep 1:26into your roster to make realistic 1:28projections based B on your team's 1:30fluctuating strengths and weaknesses 1:33trade grades are another new feature 1:34that helps managers assess the value of 1:37potential trades when managers initiate 1:39transactions with each other the AI 1:41models serve up trade insights featuring 1:44a grade for each athlete involved in the 1:46trade and a grade for the trade's 1:48overall value managers can also use the 1:51AI models to analyze structured and 1:53unstructured data to compare players 1:56estimate the potential upside and 1:57downside of starting a particular player 1:59and assess the impact of an injury to 2:02learn more and start leading your 2:03fantasy team on the journey to the 2:05playoffs check out the link in the 2:07description of this 2:08video next up generative AI chat Bots 2:12like chat gbt that are based on large 2:14language models or llms took the World 2:17by storm this year unfortunately 2:20malicious hackers moved quickly to 2:22exploit these new AI resources using 2:24chat GPT itself to polish and produce 2:26fishing 2:27emails now luckily mainstream llms have 2:31guardrails to prevent their use for 2:32scams and criminality but cyber 2:34criminals are a resourceful bunch and a 2:36range of AI tools designed specifically 2:39for malicious cyber attacks have begun 2:41to 2:42emerge right now the leading brand is 2:44called worm GPT it's an AI module based 2:47on the gptj language model and is 2:50already being used in business email 2:52compromise attacks and for other 2:54nefarious uses until now the most common 2:57way for people to identify fraudulent 2:59fishing email emails was by their 3:00suspicious wording and grammatical 3:02errors but now users can simply type in 3:05instructions for the creation of fraud 3:07emails for example they could input 3:09write an email coming from a bank that's 3:11designed to trick the recipient into 3:13giving up their login credentials the 3:16tool then produces a unique well-written 3:18and usually grammatically perfect email 3:21that's far more convincing than what 3:22most attackers would write on their own 3:25in general these generative AI tools are 3:27used for boosting three aspects of cyber 3:29text fishing researching intelligence 3:32about potential victims and writing code 3:35these tools are still in their nent 3:36stages and the threat they represent is 3:38still fairly minimal but the rise of 3:41malicious llms represents a new arms 3:43race between AI that attacks an AI that 3:46defends to learn more hit the link below 3:50thanks so much for joining me today for 3:51this episode of IBM techn if you're 3:53interested in learning more about the 3:54topics I've covered make sure you 3:56explore the links in the description of 3:57this video and again please don't forget 3:59to subscribe to our channel to stay up 4:01todate on what's going on in Tech 4:08now