Madrona Venture Group investor Aseem Datar met with about 30 startups in the past week — and all pitched AI tools.
“In the VC world, there’s a gold rush going into AI,” Datar said during a panel discussion Wednesday.
“Conversational AI Roundtable,” an event hosted at the Bellevue Club by tech consulting firm Hitech Advisors, gathered industry leaders to discuss AI’s potential impacts on enterprises, startups and society.
Business leaders from companies big and small are making plans to integrate AI into their product roadmaps as the technology continues to evolve. Presenters at Google’s I/O keynote on Wednesday said the word “AI” more than 140 times. Amazon CEO Andy Jassy highlighted generative AI in his recent shareholder letter, and Microsoft continues to push out a stream of AI product announcements.
Panelists at Wednesday’s discussion included Datar; F5 Distinguished Engineer Joel Moses; Amazon Web Services Worldwide Leader in AI/ML Denis Batalov; and Microsoft Principal Machine Learning Scientist Ameya Bhatawdekar. The conversation was moderated by Hitech Advisors CEO Haresh Sangani.
Read on for our three main takeaways.
What worries business leaders about generative AI?
- Enterprises should be aware of sharing their confidential data with AI, Datar said. If companies are training models using their internal data, the AI can pick up on their trade secrets and intellectual property. Businesses should consider safeguarding their data to maintain their competitive advantages, he said.
- Moses shared an anecdote of an executive at a major food company asking its CISO to create a system that could generate deepfakes, a form of synthetic media. The executive wanted to save time and effort when addressing employees at the company. The CISO responded by saying, “If I can put words in your mouth, so can everyone else.” Moses said these conversations happen frequently, as leaders are concerned about authentication and attribution as it relates to AI-generated content.
- Engineers may be tempted to use AI-generated computer code to boost productivity, but this can pose ethical and legal risks when it’s too similar to existing open-source code, Moses said. Batalov touted Amazon’s Code Whisperer, which offers a tool that vets AI-generated code and provides links to sources, while also checking for licensing requirements.
How can startups compete with tech giants?
- Datar believes generative AI will become “synonymous” to the current market for APIs. While companies can embed APIs for coding, fetching, and security, specialized vendors still exist to create specific tools. This trend can apply to generative AI, providing opportunities for startups to innovate in specialized verticals. He pointed to Tangibly, which uses generative AI and legal expertise to solve IP and trade secret issues.
- Startups must factor in hardware costs for running large AI models, according to Batalov. To compete with larger companies, startups will need cost-effective and powerful hardware components, he said.
- Bhatawdekar said startups should consider whether they even need to build a general-purpose AI model, which can be extremely costly. He said startups can differentiate by being strategic about their choice of model, inputted data, workflows, and distribution.
How will society be impacted by generative AI?
- While there may be displacement of workers in knowledge fields, some creative skills can be transferable to jobs in the AI industry. Moses highlighted prompt engineering, which trains AI models to deliver specific results using natural language.
- Datar said AI can help startup founders go from “0-to-1” for different aspects of running a business, including creating marketing copy, mock-ups, and software engineering. This will accelerate entrepreneurship by removing barriers to entry for non-technical people, he said. “Different vocations can now have a level playing field.”
