California's 2018 bot law covers sales and votes, not chat

The design

California's SB 1001, chaptered as Chapter 892 of the 2018 statutes, added one chapter to the state's Business and Professions Code. Signed 28 September 2018 and operative 1 July 2019, it makes it unlawful to use a bot — an automated account where substantially all actions are not a person's — to communicate online with intent to mislead someone about its artificial identity, in order to deceive them into a purchase, sale, or vote. The current codified text confirms that date and shows the duty attaches only where an operator intends deception for one of those two purposes. An operator avoids liability by disclosing the bot is a bot, clearly and conspicuously.

What the evidence says

Read against its own text, SB 1001 is narrower than it is often remembered. It creates no general duty for every chatbot to identify itself; the duty attaches only to a bot selling goods or services, or swaying an election vote, and only with intent to deceive. A companion chatbot built for conversation or entertainment sits outside that language unless it is also selling something or swaying a vote. The chapter also excludes web-hosting and internet-service providers from any duty, placing the obligation on whoever deploys the bot, not the infrastructure carrying it. This is legislative text, not a court's application of it; no litigation interpreting the intent-to-mislead element is cited here.

What it asks of people

For a covered bot, the statute asks for little beyond a disclosure a reasonable person would notice; no format, licence, or filing is specified beyond 'clear, conspicuous, and reasonably designed.' For a reader judging whether a companion product falls under the law, it asks a question the text does not resolve: whether an exchange is incentivizing a purchase, or is simply conversation. That gap matters for apps mixing free chat with paid subscriptions or in-chat purchases, where talk and a sales pitch are not distinguished by the statute.

Privacy and safeguards

SB 1001 addresses one harm — being deceived about talking to a bot — inside two contexts, sales and elections. It says nothing about data collection, emotional-design features, or minors, and created no dedicated regulator; enforcement runs through the state's general unfair-competition framework. California later adopted a separate statute aimed specifically at companion chatbots, covering a narrower population of products than SB 1001's bots. Treating SB 1001 as if it already carried that later law's duties risks crediting a 2018 disclosure statute with protections it never wrote down.

• Does a given companion product ever ask for a purchase, subscription upgrade, or vote-adjacent action inside a conversation?
• If it does, is its bot-disclosure clear and conspicuous, or left to a settings menu a user may never open?
• Does a claim about this law's reach describe SB 1001's actual text, or quietly import duties from a later, broader statute?

SB 1001 reads best as a baseline, not a companion-app rulebook: it shows California regulating bot-identity deception years before companion products became a policy focus, but drew the rule narrowly enough that most of today's companion-app design choices sit outside it.