A-Players - Hire & build high-performing teams.
A-Players #52: How ChatGPT and generative AI will disrupt recruitment, Hung Lee.
Wed Mar 08 2023
Recruiting Brain Food and Generative AI
00:03 - 11:52
- Recruiting Brain Food is a global community of recruiters and HR people that connects all the different parts of the industry together.
- Generative AI, TragedPT, GPT3 have had a huge impact on recruiting and are being used by recruiters to great effect.
- As days go by, advocates for Generative AI become more vocal while luddites become less so.
- There are risks associated with using Generative AI that need to be considered.
- Avatar AI was one of the first Generative AIs used by Hung Lee before TragedyPT became popular.
- Hung Lee experimented with Avatar AI to create a timeless profile picture for himself and found it to be very effective.
- The fidelity of results produced by Generative AIs like TragedyPT is amazing and has led to some magical moments in technology.
- The development of generative AI has been rapidly progressing, with technologies like GPT-3 and GitHub compiler paving the way.
- Chat GPT is one of the most impactful generative AI technologies, allowing for fast and high-quality text generation.
- Recruitment is a field that heavily relies on email communication, making text generation particularly valuable in this industry.
- Chat GPT can be used to write job descriptions, outreach messaging, diversity statements, and other text artifacts at a much faster rate than humans.
- Generative AI technology has achieved the miracle of high quality, high speed output at zero cost.
- The Brainfood Live event showcased various vendors demonstrating how generative AI can reduce text production costs in different areas of recruiting.
- While generative AI can improve efficiency and productivity in recruitment, there is a risk of producing generic or spam-like messages that could harm candidate engagement.
- The quality of email outreach is predicted to improve due to machine-generated messaging but may lead to an overwhelming amount of spam.
AI-generated Email Outreach and Engagement Crisis
11:24 - 17:11
- AI-generated email outreach will improve the quality of spam objectively.
- However, candidates will begin to doubt whether the message is generated by AI or a human being, leading to disengagement and an engagement crisis.
- Video is going to be more difficult to deep fake, but it will be more expensive to produce.
- The free era of sending emails is coming to an end as paywalls become more common.
- More paywalls for emails may increase trust levels since people know that the sender has paid money to message them.
- An alternative approach could be nurturing people over time instead of trying to automate and send more messages.
- All companies will look to increase their audience and cultivate an audience.
Recruiters and Generative AI
16:44 - 27:53
- Everyone needs to be a media company and cultivate an audience, including recruiters.
- Recruiters should build a following in the target community they typically recruit for.
- Generative AI won't replace recruiters, but recruiters that know how to use it will replace those who don't.
- Having skills in generative AI will increase personal productivity and capability of delivering value to employers or clients.
- CEOs should review projected headcount with providers of chat GBT or generative AI to assess potential workforce productivity gains.
- Highly skilled individuals can become recruitment marketing functions using multiple generative technologies.
- Generatively generated falsehoods are a risk due to training on false data and making inferences on true data that may still come out with falsehoods.
- Chat GPT may give one authoritative answer that sounds plausible, leading humans to trust it more than traditional search.
- Generative AI can provide authoritative answers that sound plausible, but may not be true.
- Human beings tend to believe confidence over content, making them susceptible to the confident delivery of AI-generated information.
- Access to dangerous knowledge may become easier with generative AI, which could lead to negative consequences.
- Copyright issues arise when training data is used without consent or compensation from the contributors.
- Plagiarism is a concern with generative AI as it becomes easier to copy and republish content as one's own.
- Detection software and watermarking may not be enough to prevent plagiarism in generative AI.
Live Streaming and Trust
27:37 - 34:32
- The idea of something being copied is going to be default.
- Live content that humans produce will be the default way we trust information.
- Written data and recorded video data cannot be trusted.
- In Asia, live commerce is more popular because it's harder to fake.
- M-commerce (live streaming product demos) will become more prevalent in retail.
- Live streaming can also be used for recruiting and employer branding.
- Remote work makes live streaming a challenge, but in-office folks have an advantage.
- Setting up a live stream in the office can improve employer branding.
- In-office folks have an advantage in setting up a live stream.
- Setting up a robo dog with a webcam could be entertaining but has confidentiality risks.
- Live streaming will be difficult to replicate remotely due to practical barriers.
Resources and Skills
31:31 - 34:32
- Follow recruiting brain foods and webinars for updates on GPT-3 and related tools.
- Subscribe to koreanbrayford.com for regular updates on GPT-3.
- Use GPT-3 tools regularly to improve skills and trade at a higher level in the marketplace.