Four Insights from TLV AI Week

By Tomer Y. Avni

Last week our team had the privilege to participate in the AI Week conference. AI Week is an international conference that attracts experts and leading players from academia, the private sector and government entities. At Blumberg Capital, we encounter AI-driven innovation on a daily basis. Our team often has passionate discussions about AI – its future, the perils and promises, relevant investment theses and much more. AI Week was an exceptional opportunity to hear experts’ opinions on these very issues and to meet members of the international AI ecosystem. Below are my four main takeaways from the conference:

1. AI disruption should be an interdisciplinary effort: AI disrupts traditional businesses and even legacy software, there’s no doubt about that. For this revolution to succeed, it requires a joint and interdisciplinary effort, of academia, the private sector, and governments around the world. At the conference, some government advisors shared their country’s planned strategy for the AI-era. These plans ranged on the spectrum from enabling the AI revolution (through data-sharing and supportive laws), to monitoring this revolution (through regulations that mainly deal with data privacy and explainability of models) and to exploiting this revolution for governmental use cases. The joint effort should address questions of AI ethics, data availability and privacy, smartly generating the much needed human capital for these domains, creating the vital data and computing infrastructure, algorithms security, etc. Cooperation of a variety of entities will help ensure AI benefits the most people.

2. It’s more than supervised learning: At Blumberg Capital, we meet AI-based startups almost every day. Most of them are using supervised and unsupervised learning for various use cases. AI is much more than that! Although academia has researched other types of AI for years, many of these domains have not yet been utilized by startup innovators. Emerging fields like reinforcement learning (RL) and knowledge graphs are expected to be more widely used in the coming years. Moreover, the challenge of matching the right AI branch and method to the right use case, is one of the crucial elements in building a successful AI startup. For example, RL is highly effective when building personalization and recommendation engines.

3. Data is(still) a problem: It is easy to mistakenly think that data is the easiest part of AI, after all, it’s the age of big data, isn’t it? Unfortunately, although data is collected, through various new endpoints, there are still substantial challenges in channeling, normalizing, cleansing and formatting this data for use in an AI process. First, the majority of the data is unstructured. According to research done at Accenture and mentioned by the CTO AI of the company, at the conference, 80% of the time of data engineers is spent on data preparation. We expect to see many promising startup solutions emerging in this space. Second, data privacy is another aching point for ML processes. There’s still a need for data protection, encryption, anonymization; and compliance-related enforcement engines. Third, many data lakes and warehouses should be unlocked, including government owned databases, to enable significant advancements in AI.

4. The AI value chain: Some investors at the conference expressed their belief that the AI value chain is crowded, with no room for innovation. In fact, at Blumberg Capital, we think just the opposite. AI is a dynamic, evolving driver of the new reality, the same as the cybersecurity sector was 10 years ago. As such, it’ll require many new tools and enablers to reach its full impact. We will see more innovation in the domains of observability (R&D and production), security and data (privacy, collection, enhancement and structuring). We believe many more promising AI value chain segments will emerge in the coming years.

It is an exciting time to be developing, innovating and investing with AI. While the tech continues to advance, we also need to educate consumers to build comfort with AI in their daily lives. Earlier this year, we published a survey of Americans on their thoughts and opinions on AI. The results show we’re at a consumer adoption tipping point, and while half of consumers are optimistic, half are still fearful. Tech companies have a huge opportunity to educate as they build.

We’ll continue to watch AI closely and welcome connections, feedback, deal flow, discussions and biz dev inquiries for our growing AI portfolio. We also plan to hold soon an AI event, at our Tel Aviv office, with our AI portfolio companies and other experts – so stay tuned!