The news: OpenAI’s ChatGPT, potentially the most powerful chatbot unleashed to the public, is stirring up excitement and concern.
- Based on generative AI model GPT 3.5, ChatGPT had over 1 million users within five days of its November release, per Insider.
- The chatbot’s abilities prompted Gmail creator Paul Buchheit to predict that it will completely disrupt Google Search within a year or two.
- Former Google executive Sridhar Ramaswamy said ChatGPT could disrupt Google’s business model by preventing users from clicking on links that contain ads, per Insider.
- For context, advertising generated 81% of Google parent company Alphabet’s $257.6 billion revenue in 2021, per Bloomberg.
An AI brainiac? The sensational response to ChatGPT stems from its conversational, intellectual, and creative abilities.
- ChatGPT can converse in a humanlike way, answer an array of academic and general questions, and write programming code, TV show scripts, blog posts, sheet music, and even bad poetry.
- Given the history of other chatbots spewing offensive and false content, OpenAI crafted ChatGPT with the capacity to “reject inappropriate requests,” per Insider.
- Although user reports suggest that the bot is more ready for public consumption than others, some are finding that safety controls aren’t foolproof.
- Also troubling are reports that ChatGPT sometimes provides plausible-sounding answers that are false, prompting Stack Overflow to ban it due to too many incorrect answers about programming.
What it means for Google: Chatbots like ChatGPT have the potential to provide an enhanced internet search experience beyond what’s currently available through engines like Google Search.
- Although chatbots could eventually be integral to search, don’t expect ChatGPT to pose a threat to Google any time soon.
- For starters, OpenAI has been racking up “eye-watering” compute costs since it released the system—which hasn’t yet been monetized.
- ChatGPT could be pulled from the internet while its creators figure out how to make it profitable, which could be a challenge considering its incorrect responses.
In August, Google started taking steps to reduce misinformation in Search by flagging unvetted results. Given Google’s focus on AI and ethics concerns, it’s likely cautiously exploring ways to integrate monetizable chatbots into Search.
- If ads were integrated into AI chatbots like ChatGPT, with the high compute costs, it would likely be more expensive for advertisers who have been pulling back on spending amid the economic downturn.
- Considering that 80% of consumers say customer service chatbots increased their frustration level, according to a UJET survey, per Forbes, we’ll likely see technologies like GPT 3.5 have greater influence on upgrading those tools, as opposed to search.
- With OpenAI already facing legal headwinds over generative AI, customer service chatbots that can function with smaller subsets of data for training might be lower risk for commercialization.
Dive deeper: Read on about generative AI’s transformational potential in our Trends to Watch for 2023 report.