Key Security & Privacy Actions for AI Startups

Rebecca // June 10 // 0 Comments

A Founder’s Rough Guide to Data protection strategy for a competitive edge & user trust.

Running a startup is a hustle. You’re juggling a million things at once: building your product, chasing funding, and keeping your team happy. Data privacy and responsibility can feel like just another item on the ever-growing to-do list. But prioritizing data protection from the get-go can actually be a game-changer for your startup.

Think about it. People these days are careful about their data. If you show them you take privacy seriously, it builds trust – a trust that can give you a leg up on the competition. Plus, having a solid data protection process and knowledge from the start saves you a ton of headaches down the road, especially when it comes to scaling up or partnering with bigger companies.

This post will walk you through some easy-to-implement strategies for building data protection into every stage of your startup’s life. Consider it the rough guide to the most cost-effective and impactful steps founders can take. 

These steps aren’t in the same order for all businesses. A company dealing with extremely sensitive data such as health or children’s data will need to take a more mature approach even when the company is pre-launch.  B2C companies that wish to work with large, risk-averse partners (think car manufactures or traditional banks) will need to consider audits and certification much earlier in the process.  

 

Idea Validation Stage: At this point there is an idea, and some potential founders, but not much else. You are still doing idea validation and want to know if privacy will be a blocker.  

  1. Potential founders can schedule a free 30 minute call with a privacy expert like me, who can help you weigh pros and cons of what people want, what you can do to protect data, and what will be harder or easier than you expect.  
  2. Use a password manager.  From here on out, you are a target.  (This blog post explains more https://vadimkravcenko.com/shorts/security-at-startup/ ).

 

Rush to Launch Stage: At this stage, you have a small team and not a lot of money.  The startup is moving fast, and also the product or features may be vague or changing.  The reality is that these companies will do the basics for privacy and security, but won’t have the expertise, time, or clarity to build robust protections.  In this phase, what is the minimum a cash and time-poor startup should do around data protection?

  1. The startup’s leadership (you) should emphasize to their team that they care about privacy and security.    Explicitly including it in your values goes a long way, such as occasional reminders at all-hands or in writing. You can acknowledge that you don’t have all the answers yet on how to implement privacy and security, but let it be a North Star from the beginning.  
  2. The startup should track all the privacy and security issues and questions in a secure database or tracker.  It should be easy for engineers or product teams to quickly jot down questions about data management, privacy notices, or access controls.  They may not have total product direction or resources at this point to decide about how to handle everything, but you will need this log for the next stage of growth.  
  3. Make someone responsible for privacy and security, and make sure they have some training and skills in the area.  I’ve seen CEOs and founders try to do it themselves, but  I recommend designating someone on the leadership team such as an engineering lead or CTO.  If needed, provide them some basic training on privacy and security.   Consultants like www.privacyengineer.ch provide training specifically for leadership in this position.  Alternatively, consider a Fractional Chief Privacy Officer or Fractional Responsible AI Officer (see below).  
  4. Automate and log as much as possible, including access to sensitive information.  The Cloud Security Alliance has some tips for cloud security for startups.  
  5. Create an incident management or breach plan, and check the cyber insurance policy.  The incident management plan can be a bare bones description of who to call if something goes sideways.  For example,  “call the CEO on their home number with the code ‘INCIDENT’.”   You will flesh out the incident management plan in the next stages as you build out the security defenses.  

Creating a mature data protection culture: Companies with an active customer base now need to maintain reliability and security.  This is a cultural change, as slower – but better – engineering practices and testing are implemented.  Employees may need to unlearn bad habits.  My recommendations for startups in this phase:

  1. Continue doing all of the above, including leadership’s emphasis on privacy and security. 
  2. Invest in security and privacy talent and leadership. This may be a good point for your first security hire. The non-profit First Round recommends the first security hire at 30-100 employees, which might be a good rule of thumb. If you are B2B, working in AI or handle sensitive data, you should consider separate privacy and security roles.  Fractional leadership or consultants may be cost-effective options for expertise at this point.
  3. Employee training.    Encourage all employees to use best practice and to tell security if they have any concerns. Give in-depth training to a number of cross-functional employees to understand and raise privacy and security issues.   
  4. Incorporate technical privacy and security review into the beginning of each major feature, and incorporate legal review for data protection for all major releases.
  5. Run a risk assessment: Review the issues previously tracked and prioritize the fixes.  
  6. Create a data inventory and an AI/ML inventory to map and track data and high-risk projects.  

 

Developing partnerships and preparing for audits:  The startup may begin to develop partnerships with larger companies or consider the option to be sold.  In addition to the above steps, here are some potential steps the companies will need to take.

  1. Consider security audits and certification such as ISO 27001, SOC2.  
  2. Run an external pen-test to check for security vulnerabilities.
  3. Consider a data management platform for privacy and GDPR (or local regulation) compliance.
  4. Continue to invest in employee privacy and security training on a regular basis.  

So what did I miss?  What have you seen at a small and growing company when it comes to privacy and security?

I believe in the importance of data protection for startups at every stage of growth. I’ve outlined a roadmap for founders, from the initial idea validation to developing partnerships and preparing for audits. Founders should prioritize building a culture of data protection and security, even if it means taking a slower but more secure approach. By following these steps, startups can build a strong foundation for data handling and position themselves for future success.

For AI Founders: Change the world, responsibly 

Get a startup-friendly guide to responsible AI bespoke to your startup idea and size.  Plus, get coaching from a privacy engineer.  

Proactively avoid legal issues and build a good reputation.

What is a Fractional Chief Privacy Officer?? 

Fractional CISOs and Fractional Privacy Officers can help a company improve data protection and build products that people love.  Fractional officers engage with the startup over a long term.  They advise the CEO,  run risk assessments, and develop processes and education for the company.  The startup benefits from having a “go-to” privacy person in leadership, without the costs of hiring a full-time employee.

Chief Privacy Officers are likely different from Data Protection Officers, and can help set strategy with the Chief Executive Officer on data protection.  Learn what Wikipedia has to say: https://en.wikipedia.org/wiki/Chief_privacy_officer 

 

About the Author Rebecca

Dr. Rebecca Balebako builds data protection and trust into software products. As a certified privacy professional (CIPP/E, CIPT, Fellow of Information Privacy) ex-Googler, and ex-RANDite, she has helped multiple organizations improve their responsible AI and ML programs.

Our Vision

 We work together with companies to build Responsible AI solutions that are lasting and valuable. 

Privacy by Default

respect

Quality Process

HEALTH

Inclusion

>